论文部分内容阅读
传统CORDIC算法需要通过查找表和许多乘法器才能实现矩阵的QR分解,这会导致硬件电路实现复杂,运算速度降低,此外它能够计算的角度范围也有限。针对传统CORDIC算法的缺陷,在向量模式下提出一种改进型CORDIC算法,它不需要查找表和模校正因子,只需通过简单的移位和加减运算就能实现矩阵的QR分解,从而能够减少硬件的开销,提高运算的性能,并通过重复迭代和区域变换使得该算法能够适用于所有的旋转角度。最后将该算法应用于V-BLAST接收系统的QR算法中,实现了低复杂度译码的效果。
The traditional CORDIC algorithm needs QR code decomposition of the matrix through look-up table and many multipliers, which leads to the complexity of the hardware circuit and the slowdown of the operation speed. In addition, it can calculate the angle range is limited. Aiming at the defect of traditional CORDIC algorithm, this paper proposes an improved CORDIC algorithm in vector mode. It does not need look-up table and modulo correction factor, and can only perform QR decomposition of matrix by simple shift and addition and subtraction, Reduce the hardware overhead, improve the performance of the computation, and make the algorithm suitable for all the rotation angles through repeated iterations and area transformations. Finally, the algorithm is applied to the QR algorithm of V-BLAST receiver system to achieve the effect of low complexity decoding.